Placental 11β-hydroxysteroid dehydrogenase-2 (11βHSD2) limits fetal glucocorticoid exposure and is associated with physiological stability in the premature newborn infant. Antenatal betamethasone alters 11βHSD2 activity and confers sex-specific advantages in neonatal outcome. We investigated the influence of betamethasone and sex on 11βHSD2 activity, neonatal adrenal function and clinical course in 24- to 36-wk gestation neonates from birth to day 5 of life. Univariate analyses demonstrated an interaction between timing of betamethasone exposure and sex for 11βHSD2 activity rate (P = 0.02) and umbilical arterial cortisol (P = 0.01). For infants born < 72 h following antenatal betamethasone, females had higher 11βHSD2 activity (P < 0.01) and umbilical arterial cortisol (P = 0.01) than males. Females born < 72 h of betamethasone exposure had higher day 1 urinary cortisol, if exposed to perinatal stress, than males (P < 0.01). For infants born < 72 h after betamethasone exposure, 11βHSD2 activity was negatively correlated with Clinical Illness Severity Score score (r = −0.79 P = 0.01) and positively correlated with mean arterial blood pressure (r = 0.8 P = 0.01) only in females. Sex-specific placental 11BHSD2 autoregulation following antenatal betamethasone exposure may limit adrenal suppression in females influencing physiological stability following preterm birth. A lack of adjustment in 11βHSD2 and adrenal response may contribute to the increased incidence of poor outcome observed in preterm males.
- fetal sex
preterm delivery is associated with higher neonatal mortality and long-term morbidity for male infants (11, 27). The mechanisms that confer these sex-specific differences are poorly characterized. Much attention has focused on the combination of developmental immaturity and relative adrenal insufficiency following birth. While premature infants may exhibit very low cortisol concentration without apparent ill effect (1), a lack of an adrenal response to physiological stressors may leave them vulnerable in the immediate newborn period (12).
Placental 11β-hyrdroxysteroid dehydrogenase type 2 (11βHSD2) maintains low concentrations of cortisol in the fetal circulation until late in gestation (18). The hypocortisolic fetal milieu created by placental 11βHSD2 has been suggested to be crucial for fetal hypothalamic-pituitary-adrenal (HPA) axis maturation and regulation of steroidogenesis (28). Placental 11BHSD2 activity is known to be altered by pathophysiological conditions affecting pregnancy, such as asthma (20), growth restriction (25), preeclampsia (17), and chorioamnionitis (13), common causes of preterm labor and birth.
Administration of glucocorticoids to pregnant women at risk of preterm birth decreases the incidence of neonatal respiratory distress syndrome and intraventricular hemorrhage (22). Such responses are sex-specific, with both animal (31) and human data (30) supporting a greater prophylactic effect in female infants. However, glucocorticoids exert a number of effects on the fetus and placenta. Antenatal betamethasone exposure results in a period of elevated glucocorticoid bioactivity (16) with decreased fetal cortisol levels secondary to central feedback on the fetal HPA axis (23). Betamethasone treatment is also associated with decreased placental vascular resistance (29) and vasodilatation of the placental circulation (7), and reductions in fetal growth have been demonstrated following repeated antenatal exposure (4).
Although synthetic glucocorticoids are incompletely metabolized by placental 11βHSD2 (19), alterations in 11βHSD2 activity has been demonstrated following maternal exposure (14). We hypothesized that the activity of placental 11βHSD2 would be altered following exposure to antenatal glucocorticoids in a sex-specific manner, altering fetal and neonatal adrenal function and neonatal physiological stability following preterm birth and thus contributing to the observed sex-specific differences in neonatal morbidity and mortality.
MATERIALS AND METHODS
This study was approved by the human ethics committees of the University of Newcastle and The John Hunter Hospital, Newcastle, NSW, Australia. Pregnant women presenting in preterm labor were recruited at the John Hunter Hospital as part of a prospective cohort study, according to a previously described protocol (26). Women whose infants were diagnosed with a major congenital malformation were excluded from the study. After obtaining written consent, placentae were collected within 45 min of delivery and multiple full thickness tissue blocks collected prior to storage at −80°C. Placentae and umbilical cord arterial plasma were collected from women giving birth between 24- and 36-wk gestation (n = 43).
Antenatal betamethasone results in decreased maternal and fetal cortisol levels for up to 3 days after administration (23). Therefore, preterm subjects were grouped a priori into those delivered within 72 h and those delivered > 72 h of the last maternally administered betamethasone dose. All women who proceeded to deliver preterm received at least one dose of betamethasone (11.4 mg Celestone). No participant received more than two doses of betamethasone, which were administered 24 h apart.
The following chemicals were obtained from Sigma (St. Louis, MO): cortisol (hydrocortisone), cortisone, β-nicotinamide adenine dinucleotide, bovine serum albumin fraction V, pepstatin A, benzamidine (hydrochloride hydrate), bactitracin, EDTA, DTT, and sucrose. Complete protease inhibitor cocktail tablets were obtained from Roche Diagnostics (Mannheim, Germany), Trasylol from Bayer (Leverkusen, Germany), and protein assay reagent from Bio-Rad Laboratories (Hercules, CA). [1,2,6,7-3H]-cortisol (1 mCi/ml) and biodegradable liquid scintillation cocktail were obtained from Amersham Pharmacia Biotech (Buckinghamshire, UK). Ethyl acetate, chloroform, ethanol, and chromium trioxide were obtained from BDH Laboratory Supplies (Dorset, UK). Glass-backed thin-layer chromatography plates (Adsorbosil Plus-IP containing 254 nm UV fluorescent indicators) and thin-layer chromatography tanks were from Alltech (Deerfield, IL).
11βHSD2 radiometric conversion assay.
Placental 11βHSD2 activity was measured by radiometric conversion assay as previously described (21). Briefly, pooled placental protein from each subject was homogenized in 10 volumes of 0.1 M sodium phosphate buffer (pH 7.4) containing protease inhibitors. The homogenate was centrifuged at 1,000 g for 10 min, and the supernatant centrifuged at 105,000 g for 1 h to collect the microsomal fraction in the pellet. Protein concentration was measured by the Bradford method (5). The microsomal protein was incubated at 37°C for 15 min in triplicate at three protein concentrations (100, 150, 200 μg/ml) with a saturating concentration of cold cortisol (5 μm), cofactor (1 mm β-nicotinamide adenine dinucleotide), and ∼200,000 cpm [3H]-cortisol (Amersham, Buckinghamshire, England). Following incubation, the removal of the reaction mix into 2 ml ice-cold ethyl acetate stopped the reaction and extracted the steroids into the organic phase, which was dried under high-speed vacuum.
Steroids were separated on glass-backed silica gel plates using 95:5 chloroform, methanol mobile phase. The bands were visualized under UV light, and activity was quantified using a liquid scintillation counter (model 1217 Rack Beta; LKB Wallac, Turku, Finland). 11βHSD2 activity rate was expressed as nanomoles cortisone per milligram placental protein per minute and total placental 11βHSD2 activity (activity rate × placental weight in g) was expressed as micromoles cortisone formed per hour.
Cortisol radioimmunoassay on urine and umbilical arterial blood samples.
Urine was collected on days 1–5 of life using disposable diapers (Kimberley Clark, Sydney, Australia) (10) containing a pure cellulose pad. Infants were placed on an appropriately sized diaper closed to limit evaporation. The interval for changing diapers was 4 h. Weighing the diapers before and after urine collection allowed exact calculation of 24-h urine output. Diapers were collected into a zip-lock plastic bag at room temperature for each 24-h period. Urine was extracted from the cellulose pads using a specially constructed press. Extracted urine from each 24-h period was centrifuged and stored at −80°C until required for analysis. Because of the effect of humidity in introducing error into the measured urine output, a correction for the degree and length of time in humidity was performed (2) and the corrected 24-h urine output used for all analysis.
Umbilical arterial cord blood and 24-h urinary cortisol were measured in duplicate directly by commercial radioimmunoassay (DSL Laboratories, Modesto, CA). The sensitivity of the cortisol assay was 4.7 nmol/l and the cross-reactivity of the cortisol antiserum with most other steroids was < 0.1%. Urinary cortisol results were expressed as a ratio of cortisol to urinary creatinine, determined using a commercial creatinine colormetric assay (Cayman Chemical, Ann Arbor, MI).
Perinatal distress was defined as two or more of umbilical arterial pH < 7.1, Apgar score < 5 at 5 min, or initial arterial lactate > 5 mmol/l. Physiological stability was determined by the Clinical Illness Severity Score (CRIBII) at 1 h of age in those infants ≤ 32 wk gestation (24). Cardiovascular parameters were determined at 24 h of age. Mean arterial blood pressure was determined invasively, where possible, or by oscillometry.
Demographic data are presented as median (minimum-maximum) or number (%). All results are presented as means ± SE. Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS version 14). The Student's t-test, Fisher's exact test, and ANCOVA were used for comparisons between groups, using an alpha level of 0.05. Gestational age was used as the covariate in the analysis. Post hoc correlations were conducted to assess the association between placental 11βHSD2 activity, birth weight centile, and measures of neonatal physiological stability. To increase the power of these post hoc analyses, an alpha level of 0.02 was used. Birth weight percentile was calculated using centile calculator 6.2.2 (http://www.gestation.net) with small-for-gestational age defined as birth weight < 10 percentile.
No differences in clinical characteristics or cause of preterm delivery were observed between the sexes (Table 1) or between the steroid exposure groups. For infants born within 72 h of last maternal betamethasone administration, neither the proportion exposed to two doses of betamethasone (male: 31% female: 30%) nor the time from last betamethasone exposure to birth [male: 20 (2–54) female: 24 (1–70) h] differed significantly between males and females. Similarly, in those infants born > 72 h from last maternally administered betamethasone, the length of time of delivery from steroid exposure was not significantly different between the sexes (P = 0.12).
Placental 11βHSD2 activity was positively correlated with birth weight centile (r = 0.278, P = 0.02 for activity/mg placental protein and r = 0.275, P = 0.03 for total placental activity). Total placental 11βHSD2 activity was positively correlated with gestational age (r = 0.437, P < 0.01). Small-for-gestational age infants had lower total placental 11βHSD2 activity (83.2 ± 13 vs. 194.6 ± 27.3 μmol/h, P = 0.01) than infants of normal weight for gestational age.
ANCOVA demonstrated a significant interaction between infant sex and antenatal glucocorticoid exposure in preterm infants for both placental 11βHSD2 activity rate [F(1,47) = 6.03, P = 0.02] (Fig. 1A)] and total placental 11βHSD2 activity [F(1,41) = 7.49, P < 0.01] (Fig. 1B). For infants born within 72 h of maternal glucocorticoid administration, 11βHSD2 activity rate and total placental activity was greater in females than males (activity rate: 619.2 ± 70.8 nmol·mg−1·min−1 vs. 313.4 ± 46.4 nmol·mg−1·min−1, P < 0.01, total activity: 334.2 ± 69.1 μmol/h vs. 121.3 ± 38.1 μmol/h, P < 0.01). No difference in activity rate or total activity was observed between the sexes in those infants born > 72 h after maternal betamethasone administration.
For female infants, both total placental activity and activity rate were greater in those born within 72 h compared with those born > 72 h after betamethasone exposure (total activity: 334.2 ± 69.1 μmol/h vs. 110.5 ± 18.4 μmol/h, P < 0.01, activity rate: 619.2 ± 70.8 nmol·mg−1·min−1 vs. 384.9 ± 59.4 nmol·mg−1·min−1, P = 0.02). There was no difference observed for male infants.
Fetal adrenal function.
For umbilical arterial cortisol, ANOVA demonstrated a significant interaction of fetal sex and antenatal betamethasone exposure [F(1,26) = 7.59, P = 0.01) (Fig. 2)]. For infants born within 72 h of betamethasone exposure, umbilical arterial cortisol concentration was higher in females compared with males (P = 0.01). No sex-specific difference was observed in the infants born > 72 h after maternal betamethasone administration (P = 0.14). For females, umbilical arterial cortisol concentrations were higher in those born within 72 h of betamethasone exposure than those female infants born > 72 h after betamethasone exposure (P = 0.01). No difference was observed in male infants with respect to time of delivery form steroid exposure (P = 0.16).
Neonatal adrenal function.
Umbilical arterial cortisol collected immediately following delivery was not found to correlate with day 1 24-h urinary cortisol (r = −0.186, P = 0.45). However, the relationship between placental 11BHSD2 activity rate and day 1 24-h urinary cortisol excretion was significant (r = 0.366, P < 0.05). Urinary cortisol excretion was found to change over the first week of life [F(3,45) = 3.0, P < 0.05], falling from day 3 to day 7 (P = 0.009, post hoc paired t-tests).
A significant interaction between antenatal steroid exposure, infant sex, and perinatal stress was observed for 24-h urinary cortisol excretion on day 1 of life [F(1,38) = 4.23, P = 0.04]. For infants born within 72 h of maternal betamethasone administration with evidence of perinatal stress, females had higher 24-h urinary cortisol excretion than males (P < 0.01) (Fig. 3). No difference was observed between the sexes in infants born with no indicators of perinatal stress. Similarly, no significant differences were observed in those infants born > 72 h after antenatal betamethasone exposure.
Neonatal physiological stability.
In preterm infants born < 72 h after betamethasone exposure 11βHSD2 activity was inversely correlated with CRIBII score in females (r = −0.79, P = 0.02) but not males. If delivery occurred > 72 h after betamethasone exposure, a significant inverse relationship was observed for both sexes.
For infants born within 72 h of betamethasone exposure, mean arterial blood pressure at 24 h of age was also found to significantly correlate with 11βHSD2 activity in females (r = 0.8 P = 0.01) but not males. For those infants born > 72 h after betamethasone exposure both males and females demonstrated a positive correlation between 11βHSD2 activity and mean blood pressure (female: r = 0.58, P = 0.05; male: r = 0.8, P = 0.03) relationships that neared significance for the reduced alpha level set for this analysis.
Sex is an acknowledged risk factor in the pathogenesis of common neonatal morbidities and mortality. However, the importance of sexual dimorphisms in neonatal outcome has received little investigation and remains poorly understood. We have previously demonstrated that term female fetuses appear more sensitive to changes in maternal cortisol concentration, with alterations in placental glucocorticoid receptor expression, fetal HPA axis function, and reduced growth (20). Placental 11βHSD2 has been implicated in the mechanisms contributing to these physiological changes (6). The present data supports a sexually dimorphic pattern of placental-fetal-neonatal glucocorticoid response in those infants born prematurely with placental autoregulation of 11BHSD2 activity in response to exogenous glucocorticoid exposure occurring only in females.
Maternal glucocorticoids have been demonstrated to alter expression and activity of 11BHSD2 with significant increases in placental 11BHSD2 activity observed in infants exposed to betamethasone up to 72 h before delivery (16). While the placenta only partially metabolizes betamethasone into its inactive 11-keto-derivatives (19), the acute increase in placental 11BHSD2 activity exhibited by females may have a protective action on fetal adrenal function.
A single course of antenatal betamethasone results in a period of elevated fetal glucocorticoid bioactivity (16) with decreased maternal and fetal cortisol levels due to negative feedback on the HPA axis (14). While premature infants may exhibit very low cortisol concentrations without apparent ill effect (1), a lack of response to physiological stressors may leave them vulnerable following preterm delivery (12). Suppression of fetal HPA function following antenatal betamethasone exposure may compound developmental immaturity of the adrenal limiting the ability of the preterm neonate to mount an adequate stress response and maintain physiological stability following birth.
Female sex was associated with higher umbilical arterial cord blood cortisol for those preterm infants born within 72 h of betamethasone exposure. We speculate that antenatal betamethasone treatment may have greater prophylactic effects in preterm females due to placental alterations that allow a higher level of autonomous physiological maturity compared with males. Importantly, sexual dimorphism was not due to males being delivered earlier than females, with less time for adrenal recovery following antenatal glucocorticoid exposure.
On day 1 of life, females, but not males, born < 72 h after betamethasone exposure exhibited a greater adrenal response, characterized by significantly higher 24-h urinary cortisol production, to perinatal stress. While there is no universally accepted definition of perinatal stress, both umbilical arterial acidosis defined as a cord pH < 7.1 (3) and lactic academia defined as arterial lactate > 5 mmol/l (8), have been shown to be associated with increased risk for intraventricular hemorrhage, respiratory distress syndrome, and cardiovascular morbidity (3). The manifestations of an inadequate response to critical illness are most commonly confined to the immediate newborn period. Early adrenal insufficiency associated with cardiopulmonary dysfunction (32) results in increased incidence of hypotension and need for cardiovascular support in extremely premature male infants (9). The current data support the hypothesis that the sexually dimorphic placental 11βHSD2 response to exogenous glucocorticoids exerts downstream effects on adrenal responses to physiological stressors, providing one possible explanation for the excess neonatal mortality and long-term morbidity associated with male sex (27).
11BHSD2 activity was also associated with clinical illness severity and hemodynamic stability, as evidenced by mean arterial blood pressure, in the immediate newborn period. For infants born < 72 h after maternal antenatal glucocorticoid exposure, 11βHSD2 activity was a significant predictor of neonatal physiological stability as measured by the CRIBII physiological score and mean arterial pressure in females only. Low placental 11BHSD2 activity has previously been shown to be associated with a more severe immediate clinical course, as measured by the Score for Neonatal Acute Physiology (SNAP), and a higher incidence of hypotension in infants born prematurely (15). However, SNAP is calculated from data during the first 24 h after birth, unlike CRIBII, which is calculated at 1 h of age, allowing potential treatment bias to influence the relationship with placental 11βHSD2 activity.
The current data suggest females may exhibit less HPA axis suppression following acute betamethasone exposure, enabling them to mount a more appropriate adrenal response in the immediate newborn period. Delivery > 72 h after maternal betamethasone administration allows fetal adrenal recovery with similar relationships between 11βHSD2 activity and physiological stability in both sexes, findings that confirm previous studies. The influence of fetal sex and duration of time from betamethasone exposure to birth on the relationship between placental 11βHSD2 activity and neonatal physiological stability have not previously been reported.
Human studies on expression and activity of 11BHSD2 in the placenta and fetal membranes have produced conflicting results (14, 25). The findings of previous studies have been confounded by large proportions of small-for-gestational age infants and pregnancies complicated by pathophysiological conditions known to alter 11BHSD2 activity (13, 17). However, there are a number of limitations to the current study. While a cohort of preterm infants born without antenatal steroid exposure would constitute a control population, the proven benefits of antenatal steroids for neonatal morbidity and mortality, and universal administration where possible in women presenting in preterm labor < 34 wk gestation, means that this was not feasible.
A sexually dimorphic difference in neonatal adrenal function following betamethasone exposure would be confirmed by dynamic testing of the HPA axis. While we recognize that confirmation of sex-specific differences in neonatal adrenal responses following HPA physiologic or pharmacologic challenge would support our hypothesis, there was no clinical indication to carry out these invasive procedures on this cohort of preterm infants.
Perspectives and Significance
The sexually dimorphic response to antenatal betamethasone exposure exhibited by female infants in this study may partly explain the greater prophylactic effect of this treatment in female compared with male infants (18). We have previously demonstrated sexually dimorphic differences in placental 11BHSD2 activity in response to maternal inflammation at term, conferring a female advantage. Our current findings support a sexually dimorphic difference in the preterm fetal and placental response to maternal antenatal glucocorticoid exposure, protecting the female fetus from excess glucocorticoid exposure and enabling appropriate adrenal responses to physiological stressors. With placental 11βHSD2 influencing neonatal physiological stability following preterm birth, these findings highlight potential mechanisms underlying the observed excess of male morbidity and mortality following preterm birth.
This study was funded by a John Hunter Children's Research Foundation Project Grant. M. J. Stark was supported by an Emelyn and Jennie Thomas Cardiovascular Research Scholarship and a University of Newcastle External Scholarship.
- Copyright © 2009 the American Physiological Society